Project Summary: Energy Recommendation System

LSTM-based forecasting plus recommendation logic to support sustainable grid operations.

This project combines short-term load forecasting with policy-aware recommendation rules to reduce curtailment and improve renewable integration. A modular data pipeline ingests load, weather, and calendar signals; a sequence-to-sequence LSTM delivers multi-horizon forecasts; and production serving is powered by a containerized FastAPI service.

Highlights

• Multi-horizon LSTM improves RMSE vs. ARIMA/XGBoost baselines.
• Robust under synthetic weather perturbations with clear feature attributions.
• Serving-ready: versioned artifacts, drift monitoring, and canary updates.

Explore

• Overview: Energy Recommendation System
• Case study structure: Problem, approach, architecture, operational considerations, and modeled impact.
• Repository: energy-recommendation-engine

Repo links and details available in page references.

""